wifiwave2 bridge interface "unknown" in hosts table

hap ax2, 7.11beta4 (also with previous stable).

All ethernet ports in a bridge, together with wifi1 (5ghz). Wifi client connects, but bridge hosts shows the wireless client without an interface. This makes all downstream traffic to be echoed to all bridge ports, and if one of them happens to be 100m device, the wifi download results in 70down/600up (which was the original symptom) as all tx is capped at 100mbps

Relevant config:

add admin-mac=XXXXXX ageing-time=30s arp-timeout=30s auto-mac=no comment=defconf fast-forward=no name=bridge protocol-mode=none
/interface wifiwave2
set [ find default-name=wifi1 ] channel.band=5ghz-ax .frequency=5500 .skip-dfs-channels=10min-cac .width=20/40/80mhz configuration.country=Estonia .mode=ap .ssid=red disabled=no security.authentication-types=
wpa2-psk,wpa3-psk
set [ find default-name=wifi2 ] channel.band=2ghz-ax .skip-dfs-channels=10min-cac .width=20/40mhz configuration.mode=ap .ssid=red security.authentication-types=wpa2-psk,wpa3-psk .disable-pmkid=yes .wps=disable
/interface bridge port
add bridge=bridge comment=defconf interface=ether2
add bridge=bridge comment=defconf interface=ether3
add bridge=bridge comment=defconf interface=ether4
add bridge=bridge comment=defconf interface=ether5
add bridge=bridge comment=defconf interface=wifi2
add bridge=bridge interface=ether1
add bridge=bridge interface=wifi1



With download test in progress, this is what interfaces traffic looks like:
image_2023_07_12T10_22_00_692Z.png

The screenshot shows 593.9 Mbps of download speed (that’s Tx from bridge perspective) so I don’t think this is the actual problem.

One thing that can actually mess with wireless clients (initially at least) is a long-time known “feature”: wireless interfaces (both driven by legacy driver and wifiwave2 driver) are indicated as “idle” when no wireless client is connected to that interface. Bridge then sees such interface as inactive port. When a client connects, wireless interface transitions to “active”, at this point bridge initiates xSTP procedures (and those can take between a couple of seconds to tens of seconds, depending on xSTP variety enabled) and only when these complete, bridge port gets enabled/activated. Sometimes the delay is long enough for wireless client to fail to obtain DHCP lease.
If one is 100% sure that wireless interface can not cause a loop in ethernet topology (which is mostly true), then it’s possible to deal with this issue in two ways:

  1. set wireless interface (all of them, both master and slave if there are any) with property disable-running-check=yes
    or
  2. set wireless interfaces, added to bridge as ports (under /interface/bridge/port), with property edge=yes

(I’m doing the option #1 but it’s up to you to decide which one is better in your particular use case).
In both cases bridge will not trigger xSTP procedures when first wireless client connects.

Caveat: the above described issue might not be the reason for whatever you’re seeing on your AP.

Important is to have a look at the eth2/eth3 devices, also receiving the traffic that should be meant for wifi client only.
If eth2 would have a 100mbps device connected, the speed also for the wifi client would be at 100mbps.

Bridge is configured with “none” protocol, thus no stp in play. Configuration is minimal from default (as seen above) and same thing works without issues on same routeros version but with old wifi drivers. This is the only observable difference, thus I suspect something related to it.

Right, I missed that detail. I guess you should create a supout.rif file and open issue with MT support, providing the mentioned file.

They are working on CPU side of wifiwave2 driver, they have to fix VLAN handling as well.

Repeat on this suggestion.

If nobody ever bothered to contact support with supout.rif, big change they might not know about it.

And FWIW, I don’t see this problem on my AX3… everything shows as expected, no unknowns.